TWI401944B - Noise cancellation device for an image signal processing system - Google Patents

Abstract

A noise cancellation device for an image signal processing system includes a receiving end for receiving image signals, a 3D filtering unit for adjusting a filtering parameter according to a motion estimation value, and filtering the image signals and a former filtering result for generating a current filtering result, a motion detection unit for comparing the former filtering result and the image signals, so as to generate a current motion factor and the motion estimation value according to a former motion factor, a memory unit for receiving and storing the current filter result and the current motion factor as the former filtering result and the former motion factor, and an output end for outputting the current filtering result provided by the 3D filtering unit.

Description

Noise cancellation device for video processing system

The invention relates to a noise cancellation device for a video processing system, in particular to a suitable filtering operation of a video signal according to the degree of movement of a picture, to eliminate noise, improve image quality, and maintain a static picture. Sharpness noise cancellation device.

With the rapid evolution of communication technology and computer technology, imaging applications have become more diverse. Regardless of the image application, in simple terms, the architecture can be viewed as consisting of an image source and a playback device. The image data source may be a computer, a DVD player, a wired or wireless television signal transmitting station, a video game instrument, and the like, which can output video signals through a wired (cable) or wireless (ground wave) transmission method. Signal to the playback device to display the image. In the process of signal transmission, the signal is more or less disturbed by noise or terrain, features, etc. Even when the video source or playback device processes the video signal, it may also be due to circuit or environment (such as temperature and humidity). The effect of the video signal is that the noise component that cannot be analyzed by the video signal affects the image quality output by the playback device.

Accordingly, it is a primary object of the present invention to provide a noise canceling apparatus for a video processing system.

The present invention discloses a noise cancellation apparatus for a video processing system, including a receiving end for receiving a video signal, and a three-dimensional filtering unit for adjusting a filter coefficient according to a motion estimation value, and The video is filtered with a previous filtering result to generate a current filtering result. A motion detecting unit is configured to compare the previous filtering result with the video signal received by the receiving end to generate a current according to a previous moving coefficient. a movement coefficient and the movement estimation value; a memory unit configured to receive the current filtering result output by the three-dimensional filtering unit and the current movement coefficient output by the motion detecting unit, to be stored as the previous filtering result and the a previous movement coefficient; and an output for outputting the current filtering result generated by the three-dimensional filtering unit.

Please refer to FIG. 1. FIG. 1 is a functional block diagram of a noise canceling apparatus 10 according to an embodiment of the present invention. The noise cancellation device 10 is configured to eliminate three-dimensional noise in a video processing system, and includes a receiving end 100, a three-dimensional filtering unit 102, a motion detecting unit 104, a memory unit 106, and an output terminal 108. The noise cancellation device 10 receives the video signal x(n) of the current picture through the receiving end 100, and after passing through the processing of the three-dimensional filtering unit 102, the motion detecting unit 104, and the memory unit 106, the output result is outputted by the output terminal 108(n). . In Fig. 1, y(n) represents the filtering result of the current picture, and y(n-1) represents the filtering result of the previous picture; mf(n) represents the moving coefficient of the current picture, and mf(n-1) It represents the movement coefficient of the previous picture; in addition, k represents a movement estimation value, which is defined in detail later. The moving coefficient and the motion estimation value are positively correlated with the degree of movement of the picture corresponding to the video signal x(n), that is, the more the dynamic degree of the picture is, the larger the moving coefficient mf(n) is, and the larger the k value is.

The memory unit 106 is configured to store the filtering result y(n) and the corresponding moving coefficient mf(n), and after being appropriately delayed, output to the three-dimensional filtering unit 102 and the motion detecting unit 104 when the next picture is to be received. The motion detecting unit 104 is configured to determine the dynamic degree of the current picture by comparing the filtered result y(n-1) and the video signal x(n), and generating the current according to the moving coefficient mf(n-1) of the previous picture. The movement coefficient mf(n) of the picture and the movement estimation value k. According to the motion estimation value k and the filtering result y(n-1), the three-dimensional filtering unit 102 can appropriately filter the video signal x(n) to generate a filtering result y(n); preferably, when the picture is dynamic The three-dimensional filtering unit 102 performs a two-dimensional low-pass filtering operation on the video signal x(n), and performs an infinite Impulse Response (IIR) operation on the video signal x(n) when the picture is static. In other words, the motion detection unit 104 can determine the dynamic degree of the picture, and the three-dimensional filtering unit 102 appropriately filters the video signal x(n) according to the dynamic degree of the picture. In this way, the noise canceling device 10 can effectively eliminate noise, improve image quality, and maintain the original sharpness of the static image.

In brief, the memory unit 106 stores the filtering result and the moving coefficient of the previous picture, and the three-dimensional filtering unit 102 and the motion detecting unit 104 can determine the degree of movement of the picture according to the filtering result and the moving coefficient stored by the memory unit 106. The video signal x(n) is appropriately filtered to generate a filtering result and a motion coefficient of the current picture. In this case, the noise canceling device 10 can perform appropriate filtering operations on the video signal x(n) according to the degree of movement of the screen, in addition to effectively eliminating noise, improving image quality, and maintaining the static image. The clarity. It is to be noted that the first diagram is only a functional block diagram of the noise cancellation device 10. Any circuit or device capable of implementing the three-dimensional filtering unit 102, the motion detecting unit 104, and the memory unit 106 is applicable to the present invention, and each The number, type, type, and the like of the signal connection lines between the elements are not limited to the examples shown in the drawings.

For example, please refer to FIG. 2, which is a schematic diagram of a three-dimensional filtering unit 20. The three-dimensional filtering unit 20 is configured to implement the three-dimensional filtering unit 102 in FIG. 1 , which includes a two-dimensional low-pass filter 200 , a first signal mixer 202 and a second signal mixer 204 . The two-dimensional low-pass filter 200 is coupled to the receiving end 100 for two-dimensional low-pass filtering of the video signal x(n) to output a two-dimensional low-pass filtering result x ₁ (n). The first signal mixer 202 is composed of multipliers 206 and 208 and an adder 210 for adjusting the weights of the two-dimensional low-pass filtering result x ₁ (n) and the video signal x(n) according to the motion estimation value k. To generate a mixed signal x ₂ (n) whose relationship is x ₂ (n) = k * x ₁ (n) + (1 - k) * x (n). The second signal mixer 204 is composed of multipliers 212, 214 and an adder 216 for adjusting the weights of the previous filtering result y(n-1) and the mixed signal x2(n) according to the motion estimation value k. to produce the current filtered result y (n), the relationship is y (n) = k * x 2 (n) + (1-k) * y (n-1). Therefore, when the picture is dynamic, the k value is larger, y(n) approaches x ₂ (n), and x ₂ (n) approaches x ₁ (n), that is, the filtering result y(n) tends to The two-dimensional low-pass filtering result x ₁ (n) close to the two-dimensional low-pass filter 200. Conversely, when the picture is static, the smaller the k value, the closer y(n) is to (k * x(n)+(1-k)y(n-1)), that is, the video signal x(n) The mixture with the previous filtering result y(n-1) can be regarded as the result of the infinite impulse response filtering.

Therefore, according to the dynamic degree of the picture, the three-dimensional filtering unit 20 can adjust the filtering mode to improve the quality of the dynamic picture and maintain the original definition of the static picture.

Please refer to FIG. 3 , which is a schematic diagram of a motion detection unit 30 . The motion detection unit 30 is configured to implement the motion detection unit 104 in FIG. 1 , and includes a comparison unit 300 , an operation unit 302 , a motion estimation value decision unit 304 , a low pass filter 306 , and an expander 308 . . The comparing unit 300 is configured to compare the video signal x(n) with the previous filtering result y(n-1), which is composed of a subtractor 310 and an absolute value operator 312. The comparison result output by the comparison unit 300 filters out the high frequency portion through the low pass filter 306 to make the picture more rounded, and after expanding the range by the expander 308, outputs the signal cmf to the arithmetic unit 302. It is to be noted that, in this example, the low pass filter 306 and the expander 308 enhance the accuracy of the motion coefficient mf(n) and the motion estimation value k output by the motion detecting unit 30, even if the low pass is not used. Filter 306 and expander 308 can also achieve the same purpose. The operation unit 302 can perform an operation equation on the signal cmf and the previous movement coefficient mf(n-1) to output the current movement coefficient mf(n). In this example, the operation unit 302 is composed of an attenuator 314 and a maximum value selector 316, and the corresponding operation equation is mf(n)=max(a*mf(n-1), cmf), wherein , a represents the attenuation parameter of the attenuator 314. The motion estimation value decision unit 304 outputs the motion estimation value k according to the specific rule according to the current movement coefficient mf(n).

It is to be noted that the motion detecting unit 30 of FIG. 3 is only one embodiment of the motion detecting unit 104 in FIG. 1 , and those skilled in the art can make appropriate changes without being limited to FIG. 3 . An example of this. For example, the estimated value determining unit 304 can be implemented by one of the lookup tables shown in FIG. 4; when the moving coefficient mf(n) is greater than a threshold value th2, the moving estimated value k is 1, when the moving coefficient mf(n) When it is less than a threshold value th1, the motion estimation value k is k1, and when the motion coefficient mf(n) is between the threshold value th1 and the threshold value th2, the motion estimation value k is linearity of 1 and k1. The result of the interpolation.

In a preferred embodiment of the present invention, the noise canceling apparatus 10 generates a filtering result and a moving coefficient of the current picture based on the filtering result and the moving coefficient of the previous picture. In this case, the memory unit 106 stores the filtering result and the moving coefficient of a picture. Of course, the present invention can also generate the filtering result and the moving coefficient of the current picture according to the filtering result and the moving coefficient of the previous several pictures. For example, please refer to FIG. 5, which is a functional block diagram of a noise cancellation device 50 according to an embodiment of the present invention. The noise canceling device 50 can generate the filtering result and the moving coefficient of the current picture according to the filtering result and the moving coefficient of the previous pictures. The operation mode and structure of the noise canceling device 50 are similar to those of the noise canceling apparatus 10 of FIG. 1 and will not be described again. In FIG. 5, the memory unit 506 of the noise canceling device 50 is composed of a plurality of sub-memory units, each of which can store a filtering result and a moving coefficient of a picture. In this way, the noise cancellation device 50 can generate the filtering result of the current picture according to the filtering results y(m)~y(n-1) and the moving coefficients mf(m)~mf(n-1) of the previous pictures. y(n) and the movement coefficient mf(n).

In summary, the present invention performs appropriate filtering operations on the video signal according to the degree of movement of the picture, in addition to effectively eliminating noise, improving image quality, and maintaining the original definition of the static picture.

The above are only the preferred embodiments of the present invention, and all changes and modifications made to the scope of the present invention should be within the scope of the present invention.

10, 50. . . Noise cancellation device

100. . . Receiving end

102, 20. . . Three-dimensional filtering unit

104, 30. . . Motion detection unit

106,506. . . Memory unit

108. . . Output

x(n). . . Video signal

y(n), y(m)~y(n-1). . . Filter result

Mf(n), mf(m)~mf(n-1). . . Movement coefficient

k. . . Mobile estimate

200. . . Two-dimensional low-pass filter

202. . . First signal mixer

204. . . Second signal mixer

206, 208, 212, 214. . . Multiplier

210, 216. . . Adder

300. . . Comparison unit

302. . . Arithmetic unit

304. . . Mobile estimation unit

306. . . Low pass filter

308. . . Expander

x ₁ (n). . . 2D low pass filtering results

X2(n). . . Mixed signal

Cmf. . . Signal

FIG. 1 is a functional block diagram of a noise canceling apparatus according to an embodiment of the present invention.

Figure 2 is a schematic diagram of a three-dimensional filtering unit.

Figure 3 is a schematic diagram of a motion detection unit.

Figure 4 is a schematic diagram of an estimated value decision unit.

FIG. 5 is a functional block diagram of a noise canceling apparatus according to an embodiment of the present invention.

10. . . Noise cancellation device

100. . . Receiving end

102. . . Three-dimensional filtering unit

104. . . Motion detection unit

106. . . Memory unit

108. . . Output

x(n). . . Video signal

y(n), y(n-1). . . Filter result

Mf(n), mf(n-1). . . Movement coefficient

k. . . Mobile estimate

Claims (19)

A noise cancellation device for a video processing system includes: a receiving end for receiving a video signal; and a three-dimensional filtering unit for adjusting a filter coefficient according to a motion estimation value, and the video signal The signal is filtered with a previous filtering result to generate a current filtering result; a motion detecting unit is configured to compare the previous filtering result with the video signal received by the receiving end to generate a current according to a previous moving coefficient a movement coefficient and the movement estimation value; a memory unit configured to receive the current filtering result output by the three-dimensional filtering unit and the current movement coefficient output by the motion detecting unit, to be stored as the previous filtering result and the a previous movement coefficient; and an output for outputting the current filtering result generated by the three-dimensional filtering unit. The noise cancellation device of claim 1, wherein the three-dimensional filtering unit comprises: a two-dimensional low-pass filter coupled to the receiving end for performing two-dimensional low-pass filtering on the video signal to output a two-dimensional low-pass filtering result; a first signal mixer coupled to the receiving end and the two-dimensional low-pass filter for adjusting the two-dimensional low-pass filtering result and the video according to the motion estimation value The signal is weighted to generate a mixed signal; and a second signal mixer is coupled to the memory unit and the first signal mixer for adjusting the previous filtering result and the mixed signal according to the motion estimation value The weights are generated to produce the current filtered result. The noise cancellation device of claim 2, wherein the first signal mixer increases the weight of the two-dimensional low-pass filtering result when the motion estimation value is large, and the motion estimation value is smaller. At the same time, the weight of the two-dimensional low-pass filtering result is reduced. The noise cancellation device of claim 2, wherein the first signal mixer comprises: a first multiplier coupled to the two-dimensional low-pass filter for calculating the two-dimensional low-pass filtering result and a multiplication result of the motion estimation value to generate a first multiplication result; a second multiplier coupled to the receiving end, configured to calculate a multiplication result of the complement of the video signal and the motion estimation value, Generating a second multiplication result; and an adder coupled to the first multiplier and the second multiplier for accumulating the first multiplication result and the second multiplication result to generate the mixed signal. The noise cancellation device of claim 4, wherein the complement of the motion estimation value is a difference between 1 and the motion estimation value. The noise cancellation device of claim 2, wherein the second signal mixer increases the weight of the mixed signal when the motion estimation value is larger, and reduces the weight when the motion estimation value is smaller. The weight of the mixed signal. The noise cancellation device of claim 2, wherein the second signal mixer comprises: a first multiplier coupled to the first signal mixer for calculating the mixed signal and the motion estimation value. a multiplication result to generate a first multiplication result; a second multiplier coupled to the memory unit for calculating a multiplication result of the complement of the video signal and the motion estimation value to generate a second multiplication method And an adder coupled to the first multiplier and the second multiplier for accumulating the first multiplication result and the second multiplication result to generate the current filtering result. The noise cancellation device of claim 7, wherein the complement of the motion estimation value is a difference between 1 and the motion estimation value. The noise cancellation device of claim 1, wherein the dynamic detection unit comprises: a comparison unit coupled to the receiving end and the memory unit for comparing the video signal with the memory unit The result of the previous filtering is used to output a comparison result; an operation unit is coupled to the comparison unit and the memory unit for performing the comparison result output by the comparison unit according to the previous movement coefficient stored by the memory unit An operation equation for outputting the current movement coefficient; and a movement estimation value determining unit coupled to the operation unit for outputting the movement estimation value according to the current movement coefficient. The noise cancellation device of claim 9, wherein the comparison unit comprises: a subtractor coupled to the receiving end and the memory unit for calculating the video signal and the previous filtering stored by the memory unit a difference between the results to produce a subtraction result; and an absolute value operator coupled between the subtractor and the arithmetic unit for calculating an absolute value of the subtraction result to generate the comparison result. The noise cancellation device of claim 9, wherein the operation unit comprises: an attenuator coupled to the memory unit for attenuating the previous filtering result stored by the memory unit to generate an attenuation result; And a maximum value selector coupled to the attenuator and the comparing unit for obtaining the attenuation result and a maximum value of the comparison result to generate the current movement coefficient. The noise cancellation device of claim 9, wherein the motion estimation value determining unit is a lookup table. The noise cancellation device of claim 9, further comprising a low pass filter coupled between the comparison unit and the operation unit for calculating a low pass filtering result of the comparison result. The noise cancellation device of claim 9, further comprising an expander coupled between the comparison unit and the operation unit for expanding a range of the comparison result. The noise cancellation device of claim 1, wherein the memory unit stores a filtering result and a motion coefficient of a picture. The noise cancellation device of claim 1, wherein the memory unit stores the filtering result and the moving coefficient of the plurality of pictures. The noise cancellation device of claim 16, wherein the memory unit comprises: a signal receiving end for receiving the current filtering result and the current moving coefficient; and a signal output end for outputting the previous filtering result and The previous movement coefficient; and a sequence of sub-memory units coupled between the signal receiving end and the signal output end, each sub-memory unit of the serial sub-memory unit is configured to store a filtering result and a moving coefficient of a picture. The noise cancellation device of claim 1, wherein the motion estimation value is positively correlated with a degree of movement of a picture corresponding to the video signal. The noise cancellation device of claim 1, wherein the movement coefficient is positively correlated with a degree of movement of a picture corresponding to the video signal.